In the increasingly competitive global market environment, companies face multiple challenges in production decision-making, including resource optimization, cost control, and uncertainty management. Traditional production decision-making methods rely on empirical judgment or simplified models, which struggle to handle the dynamism and randomness in complex production processes. Existing optimization algorithms also have limitations such as overly simplified models, high computational complexity, and inadequate handling of uncertain factors. To address these issues, this paper proposes an integrated decision framework that combines multi-stage dynamic programming, Monte Carlo simulation, and greedy algorithms. By dividing production stages through dynamic programming and establishing a global optimization model, Monte Carlo simulation quantifies the impact of random factors, while the greedy algorithm quickly solves local optimal strategies to reduce computational complexity. Experiments show that this method can effectively balance inspection, disassembly, and replacement costs in scenarios involving component assembly and multi-process semi-finished product production. Additionally, this paper reduces resource waste through a pre-interception mechanism for defect rates, enhancing the company's adaptability to market uncertainties. The research provides a decision tool with both efficiency and precision for multi-objective optimization in complex production systems, helping companies achieve dual goals of minimizing costs and maximizing profits in dynamic environments.
Huang et al. (Mon,) studied this question.